[detection-analysis] Detection Analysis Report — 2026-07-11 (last 24h)

[github-actions[bot]]

Detection Analysis Report — Last 24h

Summary

• Window evaluated: 2026-07-10T13:10:49Z → 2026-07-11T13:10:49Z (last 24 full hours, UTC)
• Total runs analyzed: 103
• Detection-enabled runs: 28 (27.2%)
• Regular runs: 75 (72.8%)
• Misconfigured workflows found: 4

Warning

4 workflows show detection misconfigurations in this window. See the Misconfigured Workflows section for recommended fixes.

Detection-enabled runs outperformed regular runs on reliability this window: 92.9% success vs 88.0%, at a modest token premium (avg 43,901 vs 39,252 tokens/run).

Comparison Chart

Metric	Regular Runs	Detection Runs
Total runs	75	28
Success rate	88.0%	92.9%
Avg tokens	39,252	43,901
Failure count	9	2
Misconfigured count	—	4

Misconfigured Workflows

Workflow	Misconfiguration type	Runs	Recommended fix
Smoke CI	Detection explicitly disabled on active workflow (6 runs)	6	Set gh-aw-detection: true in frontmatter — workflow is active (>3 runs) and should have detection enabled. (§29146158533)
Copilot Agent Prompt Clustering Analysis	Detection absent on audit/analysis/report workflow	1	Add gh-aw-detection: true — the name (...Analysis) indicates an analysis workflow that should run detection. (§29148804074)
daily-experiment-report	Detection-enabled run failed	1	Investigate the detection job/step failure; check agent logs for detection error patterns. (§29146130589)
Daily Max Ai Credits Test	Detection-enabled run failed	1	Investigate the detection job/step failure; check agent logs for detection error patterns. (§29151074694)

View All Run Metrics

Per-workflow breakdown (24h window). Detection column: ✅ all = every run detection-enabled, — = none, n/N = mixed.

Workflow	Detection	Runs	Success rate	Avg tokens
Design Decision Gate 🏗️	—	9	89%	11,433
Impeccable Skills Reviewer	—	9	100%	6,757
Test Quality Sentinel	✅ all	9	100%	7,809
Matt Pocock Skills Reviewer	—	9	100%	13,728
PR Sous Chef	—	9	78%	32,357
PR Code Quality Reviewer	—	8	100%	14,918
AI Moderator	✅ all	7	100%	104,120
Smoke CI	—	6	0%	0
PR Description Updater	—	6	100%	10,096
Issue Monster	—	5	100%	21,705
Avenger	✅ all	2	100%	0
Daily Rendering Scripts Verifier	✅ all	1	100%	32,337
Weekly Editors Health Check	—	1	100%	10,630
daily-experiment-report	✅ all	1	0%	82,385
Release	—	1	100%	9,632
Stale PR Cleanup	—	1	100%	5,859
Smoke Copilot	—	1	100%	1,036,186
Instructions Janitor	—	1	100%	23,133
Terminal Stylist	—	1	100%	122,456
ESLint Miner	—	1	100%	12,427
Dev	—	1	100%	226,482
Contribution Check	✅ all	1	100%	27,430
Package Specification Extractor	—	1	100%	168,780
Copilot Agent Prompt Clustering Analysis	—	1	100%	29,617
Daily Documentation Updater	✅ all	1	100%	100,276
GitHub API Consumption Report Agent	✅ all	1	100%	42,050
Draft PR Cleanup	—	1	100%	6,437
Daily Max Ai Credits Test	✅ all	1	0%	2,852
Constraint Solving — Problem of the Day	✅ all	1	100%	6,481
Package Specification Enforcer	—	1	100%	59,043
Team Status	—	1	100%	33,104
Daily Go Function Namer	✅ all	1	100%	14,321
PR Triage Agent	—	1	100%	90,812
Claude Code User Documentation Review	✅ all	1	100%	89,317
Auto-Triage Issues	✅ all	1	100%	5,073

View Historical Trend

Detection adoption and success rate over the last 30 days (15 days of history available).

Detection run volume has grown steadily (12→37/day range), and detection success rate has recovered to the ~90% band after a dip on 2026-07-09.

Recommendations

1. Smoke CI — enable detection or confirm intent. Six runs this window with detection explicitly off and 0% success/0 tokens suggest a non-agentic smoke job; if it is agentic, set gh-aw-detection: true. If it is a pure CI smoke test, the 0% is expected and detection is legitimately N/A — document that to suppress future flags.
2. Copilot Agent Prompt Clustering Analysis — add gh-aw-detection: true. Its name marks it as an analysis workflow; detection should be on for audit/analysis/report/monitor workflows.
3. Investigate the two failed detection runs (daily-experiment-report, Daily Max Ai Credits Test). Both are single detection-enabled runs that failed — pull their agent logs and confirm whether the failure originates in the detection step or elsewhere.
4. Keep detection enabled broadly. In this window detection runs had a higher success rate (92.9% vs 88.0%) for only ~12% more tokens — the reliability gain justifies the cost.

References:

• §29146158533 — Smoke CI
• §29148804074 — Copilot Agent Prompt Clustering Analysis
• §29146130589 — daily-experiment-report

Warning

Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

• awmgmcpg

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "awmgmcpg"

See Network Configuration for more information.

Generated by 🔎 Detection Analysis Report · 180.1 AIC · ⌖ 35.6 AIC · ⊞ 7.1K · ◷

• expires on Jul 14, 2026, 5:16 AM UTC-08:00

[github-actions[bot]]

This discussion has been marked as outdated by Detection Analysis Report.

A newer discussion is available at Discussion #45089.